Urban Digital Twins (UDTs) rely on Internet of Things (IoT) sensors as fundamental enablers for bridging the physical and virtual domains, but face substantial challenges in data integration, as well as in interoperability. Current sensor data registration to metadata catalogs relies on time-consuming, human-initiated processes susceptible to errors, resulting in stale entries when services evolve. Urban IoT products with hundreds or thousands of individual sensors require substantive grouping and highquality metadata to enable discoverability across different user groups like municipal agencies, city planners, and transportation departments. This thesis introduces WRENCH, an open-source modular end-to-end framework that enables continuous automated registration of IoT sensor data into metadata catalogs utilizing Harvesters, Groupers, MetadataEnrichers, and Catalogers as components. Continuous harvesting is achieved through scheduled pipeline runs and component state management. The key innovation of the framework is KINETIC, a novel clustering algorithm designed explicitly for grouping IoT devices that leverages keyword extraction, co-occurrence network analysis, and community detection in identifying thematic groups of devices. Unlike conventional topic modeling techniques that fail with heterogeneous, unbalanced sensor data, KINETIC can identify underrepresented clusters without compromising semantic coherence. Evaluation on the Hamburg, Osnabrück and Munich datasets demonstrates KINETIC’s superior performance compared to traditional methods like LDA, with significantly better clustering quality metrics (Normalized Mutual Information, Homogeneity, Completeness, and V-Measure). Large Language Model features are built into the system to automatically create detailed, descriptive titles and descriptions that manage vocabulary differences between stakeholder groups. WRENCH supports multilingual data processing, provides template pipelines for rapid deployment, and provides ongoing automated updates to keep current metadata catalog entries. The framework addresses central challenges in urban IoT metadata management and facilitates the growing utilization of digital twinning technologies. Released as an open-source framework, WRENCH provides a foundation for scalable automation across heterogeneous sensor ecosystems and facilitates further developments in Urban Digital Twin implementations.     «

Urban Digital Twins (UDTs) rely on Internet of Things (IoT) sensors as fundamental enablers for bridging the physical and virtual domains, but face substantial challenges in data integration, as well as in interoperability. Current sensor data registration to metadata catalogs relies on time-consuming, human-initiated processes susceptible to errors, resulting in stale entries when services evolve. Urban IoT products with hundreds or thousands of individual sensors require substantive grouping a...     »